hkaiser changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar.cct.lsu.edu | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | Buildbot: http://rostam.cct.lsu.edu/ | Log: http://irclog.cct.lsu.edu/
eschnett has joined #ste||ar
hkaiser has quit [Quit: bye]
bobakk3r has joined #ste||ar
bobakk3r has left #ste||ar [#ste||ar]
eschnett has quit [Quit: eschnett]
nikunj has joined #ste||ar
nikunj has quit [Ping timeout: 255 seconds]
nikunj has joined #ste||ar
nikunj has quit [Ping timeout: 245 seconds]
nikunj has joined #ste||ar
nikunj has quit [Quit: Leaving]
<heller>
jbjnr__: I guess it is best if we chat about it tomorrow morning or so
<heller>
jbjnr__: I was not aware of it being an issue
<heller>
the barrier seemed to have worked without problems so far..
<jbjnr__>
heller: yes. I think I found the real problem
<jbjnr__>
I have a condition where I yield_while - but it happens inside the backgound work task, so once that happens, no progress is made it seems
<jbjnr__>
(just curious? is it possible to know if you're inside a background work task?)
<jbjnr__>
hkaiser: heller I think I fixed the hang problem and have relaunched jobs on daint. Fingers crossed.
<jbjnr__>
(side effect is that it runs slightly faster now too)
eschnett has joined #ste||ar
<hkaiser>
jbjnr__: what was the problem?
<jbjnr__>
hkaiser: might not be fixed after all. I had my first deadlock just now. I was calling yield_while on the background thread. once all the background tasks have yielded, there is no work being done any more.
<jbjnr__>
hence the deadlock
<jbjnr__>
I've replaced the yield, but there are others in the code that might cause trouble.